FastMCP Tools Calculator

# MCP MuleSoft Assistant The **MCP MuleSoft Assistant** is a lightweight Python package that provides a set of ready‑to‑use tools for scraping MuleSoft documentation. It can retrieve runtime versions, DataWeave compatibility tables, and connector version information – all without leaving your development environment. --- ## 📦 Installation The project uses a standard `requirements.txt` file. From the repository root run the commands that match your operating system. ### Windows ```cmd python -m venv .venv # (optional) create an isolated environment .venv\Scripts\activate # activate the venv pip install -r requirements.txt # install dependencies ``` ### Linux / macOS ```bash python3 -m venv .venv # (optional) create an isolated environment source .venv/bin/activate # activate the venv pip install -r requirements.txt # install dependencies ``` If you want to use the assistant as an importable package in other projects you can install it in *editable* mode (works on all platforms): ```bash pip install -e . ``` --- ## 🚀 Usage All public functions live in `mulesoft_assistant.mulesoft_server`. A typical workflow looks like this: Direct function call(s) ```python from mulesoft_assistant import mulesoft_server as mms # Get every EDGE and LTS runtime version that MuleSoft publishes all_versions = mms.get_mulesoft_runtime_versions() print("All runtime versions:", all_versions) # Get only the latest EDGE and LTS releases latest = mms.get_latest_mulesoft_versions() print("Latest runtimes:", latest) # DataWeave compatibility matrix dw = mms.get_dataweave_versions() print("DataWeave versions:", dw) # Connector version compatibility connectors = mms.get_connector_versions() print("Connector versions:", connectors) ``` If you prefer an object‑oriented wrapper, you can create a tiny client class that forwards calls to the server module. This makes it easy to inject the client into other services or to mock it in tests. ```python """Simple client wrapper around the MCP MuleSoft Assistant functions.""" from mulesoft_assistant import mulesoft_server as mms class MuleSoftAssistantClient: """Encapsulates the public API of the assistant. The methods simply delegate to the underlying module functions, but the class provides a clear place to add caching, logging, or custom error handling in the future. """ def get_all_runtimes(self): return mms.get_mulesoft_runtime_versions() def get_latest_runtimes(self): return mms.get_latest_mulesoft_versions() def get_dataweave_matrix(self): return mms.get_dataweave_versions() def get_connector_matrix(self): return mms.get_connector_versions() # Example usage if __name__ == "__main__": client = MuleSoftAssistantClient() print("All runtimes:", client.get_all_runtimes()) print("Latest runtimes:", client.get_latest_runtimes()) print("DataWeave matrix:", client.get_dataweave_matrix()) print("Connector matrix:", client.get_connector_matrix()) ``` ### FastMCP Client Integration You can also interact with the MuleSoft assistant programmatically using the FastMCP client, as described in the [FastMCP documentation](https://gofastmcp.com/getting-started/quickstart#call-your-server): ```python import asyncio from fastmcp import Client # Connect to the assistant running as an HTTP server client = Client("http://localhost:8000/mcp") async def get_latest_runtimes(): async with client: # Call the get_latest_mulesoft_versions tool result = await client.call_tool("get_latest_mulesoft_versions", {}) print("Latest runtimes:", result) asyncio.run(get_latest_runtimes()) ``` Note that: - FastMCP clients are asynchronous, so we use `asyncio.run` to run the client - We must enter a client context (`async with client:`) before using the client - You can make multiple client calls within the same context ### Client Integration The MuleSoft assistant can be easily integrated with various MCP clients using the `fastmcp install` command. Here are the supported integration options: ```bash fastmcp install --help Usage: fastmcp install COMMAND Install MCP servers in various clients and formats. ╭─ Commands ───────────────────────────────────────────────────────────────────────╮ │ claude-code Install an MCP server in Claude Code. │ │ claude-desktop Install an MCP server in Claude Desktop. │ │ cursor Install an MCP server in Cursor. │ │ gemini-cli Install an MCP server in Gemini CLI. │ │ mcp-json Generate MCP configuration JSON for manual installation. │ ╰────────────────────────────────────────────────────────────────────────────────── ``` #### Direct Client Integrations For supported clients, you can directly install the assistant using the specific command: - **Claude Desktop**: `fastmcp install claude-desktop python mulesoft_assistant\mulesoft_server.py` - **Claude Code**: `fastmcp install claude-code python mulesoft_assistant\mulesoft_server.py` - **Cursor**: `fastmcp install cursor python mulesoft_assistant\mulesoft_server.py` - **Gemini CLI**: `fastmcp install gemini-cli python mulesoft_assistant\mulesoft_server.py` #### Manual Configuration (mcp-json) For other MCP-compatible clients or manual setup, use the mcp-json generator: ```bash fastmcp install mcp-json --server-spec python mulesoft_assistant\mulesoft_server.py --copy ``` This command generates a JSON configuration that can be used with any MCP-compatible client. The `--copy` flag copies the configuration to your clipboard for easy pasting into your client's configuration file. --- ## ✨ Features & Tools --- ## ✨ Features & Tools | Feature | Description | Primary Function | |---------|-------------|------------------| | **Runtime discovery** | Scrapes the MuleSoft runtime download page and extracts both EDGE and LTS versions. | `get_mulesoft_runtime_versions()` | | **Latest runtime shortcut** | Returns only the most recent EDGE and LTS releases for quick reference. | `get_latest_mulesoft_versions()` | | **DataWeave compatibility** | Parses the DataWeave version matrix to show which runtime supports which DW version. | `get_dataweave_versions()` | | **Connector version matrix** | Retrieves connector‑to‑runtime compatibility tables for all official MuleSoft connectors. | `get_connector_versions()` | ### Prompts & MCP Integration The assistant is designed to be used from **MCP (Model‑Context‑Protocol) tools**. Example prompts you can feed to an MCP‑enabled LLM: * "_Give me the latest MuleSoft LTS runtime version._" * "_Which DataWeave version works with Mule 4.4.0?_" * "_List all connector versions compatible with Mule 4.5.0 EDGE._" When these prompts are routed through an MCP tool that calls the corresponding Python function, the LLM receives a structured answer instead of raw scraped HTML. --- ## 📂 File Structure ``` mulesoft_assistant/ # Core package ├── __init__.py # Makes the folder a package ├── mulesoft_server.py # Public API – scraping tools ├── mulesoft_constants.py # URLs, headers, and static config ├── mulesoft_utils.py # Helper utilities (parsing, validation) └── http_client.py # HTTP client utilities README.md # This documentation file requirements.txt # Python dependencies ``` --- ## 🛠 Naming Conventions (reminder) * **Folders** – lower‑case with underscores, prefixed with `mcp_` when they belong to the MCP ecosystem. * **Python files** – lower‑case with underscores; use a descriptive suffix (`_constants`, `_utils`, `_server`). * **Functions / variables** – `snake_case`; verbs for actions (`get_`, `parse_`, `extract_`). * **Constants** – `UPPER_SNAKE_CASE` and kept in dedicated constant modules. --- ## 📄 License This project is licensed under the MIT License – see the `LICENSE` file for details.